VM delete pre-save was calling save_vm_instance() without template_id/
template_name, which unconditionally overwrites those fields on the
existing cached row - so any non-force delete attempt (even one that
later failed) permanently detached the VM from its source template and
broke reinstall. Now the previous values are carried through.
Also extend reinstall eligibility and the reinstall worker to cover
QEMU VMs created via "Create VM from ISO" (template_name holds the ISO
volid, same storage:content/file pattern already handled for LXC CT
templates), recreating the VM from the same ISO on reinstall.
A guest was implicitly limited to a single allocation: lookups were
"(server, vmid) -> one row", release_ip_by_vmid freed one address, and
the link tool skipped guests that already had one. An instance can now
hold a primary address plus any number of aliases, managed from the
Network tab of the instance and visible on the IPAM pages.
Schema and service:
- ipam_allocations gains is_primary, assignment_kind, target_interface
and apply_* columns; migration 43 marks existing rows as primary
- find_allocations_by_resource/set_primary; release_ip_by_vmid now frees
every address of a deleted guest instead of just the first one
- next-available search may scan the whole subnet when a network has no
auto-assign pool, skipping node and cached guest addresses so it never
hands out an address that is already live
Applying to the guest (services/guest_ip_service.py):
- reuses script_engine, so LXC goes through pct exec and VMs through the
guest agent; the shell script synchronises the whole alias list of an
interface, which makes add, remove and retry the same code path
- persists via ifupdown, netplan, NetworkManager or sysconfig and never
restarts guest networking; an unknown stack degrades to runtime_only
and an unreachable guest keeps the allocation as pending
API and UI:
- /ipam/api/guests/{server_id}/{vmid}/addresses: list, add, apply,
primary, delete; allocations list accepts guest filters
- instance list exposes ips[] alongside the primary ip
- addresses card on the Network tab, "+N" badge in the IP column,
primary badge in the allocations table
Fixes found along the way:
- link-allocations registers every address of a guest, carries the MAC
over, refuses to take an address away from another live guest and no
longer matches allocations by resource name (which leaked an address
between same-named guests on different servers)
- IPAM history no longer fails to serialise datetime values, which used
to abort re-linking with a 500
Opening an LXC console left a failed task on the Proxmox node:
failed waiting for client: timed out
TASK ERROR: command '/usr/bin/termproxy 5900 ... vzctlconsole101' failed
connectTerminal() called GET /api/{sid}/container/{vid}/terminal first,
which creates a termproxy session on the node, but only the `node` field
of the response was used - port and ticket were discarded. The ws
endpoint then created a second termproxy session and connected to that
one. Nobody ever attached to the first, so Proxmox waited ~10s for a
client and failed the task.
The console also rendered a black screen in dev: React StrictMode mounts
the effect twice, connect* are async, and the effect cleanup ran before
they had assigned wsRef/termRef, so it had nothing to clean up. Two
xterm instances ended up in the same container - the first one filled
the viewport while the live one was pushed out of the overflow-hidden
area, which is why the badge read "connected" over an empty screen.
- drop the redundant REST call; the ws endpoint owns the termproxy
- remove the now unused GET /api/{server_id}/container/{vmid}/terminal
- add a per-session cancellation guard: a stale connect closes its own
socket and disposes its own terminal without touching refs that may
already point at the live session
- route mount and the reconnect button through a single startSession()
that cancels and tears down the previous session first
- apply the same guard to the node shell page
Disk move was hardcoded to the QEMU move_disk Proxmox API and was
never reachable in the UI for LXC (no Hardware tab for containers,
and the disk list regex only matched VM device names).
- add move_container_disk() using lxc/{vmid}/move_volume
- add POST /api/{server_id}/container/{vmid}/disk/move endpoint
- route useMoveDisk by instance type (vm vs container prefix)
- recognize rootfs/mp\d+ as LXC disk devices
- add a Move disk card on the Settings tab for LXC instances
- filter target storage list by supported content type
(rootdir for LXC, images for VM) instead of showing all storages
Fix nested h-screen flexbox layout bug that let the noVNC canvas overflow
below the viewport until the browser zoom level was changed (missing
min-h-0 on flex children). Also add a fuller console control bar: power
menu (reboot/shutdown/hard reset/hard stop), extra-keys menu (Ctrl+Alt+Del,
Ctrl+Alt+Backspace, Tab, Esc, Ctrl+Alt+F1-F12), scale toggle, clipboard
paste to guest, and console screenshot. Adds a VM hard-reset backend
action (qemu status/reset) alongside the existing start/stop/restart/shutdown.
Derive a human-readable OS name (e.g. Debian 13, Ubuntu 24.04) from the
vztmpl volid instead of showing the raw template path like
local:vztmpl/debian-13-standard_13.1-2_amd64.tar.zst in the instances list.
Applied at the display layer, so existing instances are fixed too.
Remove the previous owner's SSH keys and apply the new owner's keys
(profile + library) when ownership is reassigned. If the new owner has
no keys, the old keys are simply removed.
QEMU VMs: updates cloud-init sshkeys via Proxmox API.
LXC containers: writes /root/.ssh/authorized_keys via exec.
The swap is best-effort — failures are logged but do not block the
ownership change.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Images: show real ISO files already on the node storage (incl. those
present before the node joined the panel) via existing node/isos endpoint
- ISO dialog: rename "detach" to "Eject", show current image inside the
select (drop overflowing side text), add "boot from disk" checkbox
- Backend: on eject, put disk first in boot order; on mount, optional
"boot from this ISO" puts CD-ROM first
- Reboot to apply the new boot order via hybrid restart: graceful ACPI
shutdown with forceStop fallback, then start (fixes guest-ping timeout
on freshly installed OS without QEMU Guest Agent)
- Update iso_next_steps hint to reflect the automated flow
- Invalidate workspace list cache after assigning servers/users so
server/user counts on workspace cards update immediately.
- Invalidate node/VM caches after workspace server/user assignment so
the Nodes page reflects visibility changes right away.
- Soft-delete the cached VM/LXC row and broadcast vm_deleted the moment
a VM/container is destroyed, instead of waiting for the 10s
background sync_vm_cache tick.
- Invalidate the VM list on the frontend when a deploy task completes,
since the backend already persists the new instance to cache
synchronously before marking the task done.
/api/cluster/topology and /api/resources/all read every ProxmoxServer row
regardless of the caller access or the X-Active-Workspace header, so the
Cluster page listed all servers and dashboard aggregates counted the whole
panel instead of the active workspace. Extract the access + workspace filter
out of /api/servers into get_visible_server_ids() and apply it in all three.
Live metrics were also gathered server by server with the cache lock held
across the network calls, so a single unreachable cluster stalled the request
(21-31s in the logs) and concurrent requests queued behind it. Fetch in
parallel, skip servers already known offline, drop the redundant
is_connected() probe, and hold the lock only around the cache dict.
/api/resources/all now orders by server id: without it Postgres returned heap
order, which shifts whenever update_status rewrites a row, so the server list
reshuffled between requests.
App Store install/update pipeline reliability fixes:
- explicit nameserver (APPSTORE_NAMESERVER) on App Store LXCs and the golden
template build CT, instead of inheriting a node resolver that may be
unreachable from the container network (e.g. Tailscale-only DNS)
- golden template storage/resolution moved from one global setting to
per-server_id (migration 37: app_operations.server_id) — building it on one
Proxmox server no longer overwrites the template configured for another
- golden template automated build fixed: vzdump export step no longer fails
on "unable to parse directory volume name 'vztmpl/probe'"
- bind-mount host directories are pre-created with permissive ownership
before `docker compose up`, fixing Permission denied for non-root
containers (e.g. MariaDB) writing to Docker-auto-created root-owned dirs
- delete/teardown now waits for Proxmox to actually confirm the destroy task
finished (was silently proceeding after a short timeout, letting a new
install reuse the VMID while the node was still wiping the old disk)
- pct create retries a few times on transient "got lock" config errors
- failed pipeline steps now report the tail of command output (4000 chars)
instead of the head, so the real error isn't buried under pull/progress noise
- install error box in the UI wraps long text and scrolls instead of
stretching the dialog
Also carries forward in-progress script catalog / instances / RBAC changes
that were already uncommitted in the working tree before this session.
Bump to v1.9.1, update WIKI.md/docs/CHANGELOG accordingly.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
1. vms.py: catch-all маршруты /vm/{vmid}/{action} и /container/{vmid}/{action}
перенесены в конец файла — раньше они перехватывали POST /unlock,
/cloud-init и /exec, и эти операции всегда падали с 400 "Invalid action".
2. task_queue_service: зависшие после рестарта running-задачи помечаются
failed при старте — иначе очередь bulk-операций блокировалась навсегда.
3. api-client.ts: 401 от /api/auth/login больше не вызывает перезагрузку
страницы — пользователь видит сообщение "неверный логин/пароль".
4. task_queue_service: bulk shutdown теперь вызывает shutdown_vm/
shutdown_container (graceful) вместо жёсткого stop.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- Сертификаты: методы клиента + endpoints GET/POST/DELETE
/nodes/{node}/certificates, хуки и вкладка «Сертификаты» в
NodeAdminManager (просмотр, загрузка PEM cert+key, удаление кастомного,
бейдж срока действия)
- Bulk migrate: тип bulk_migrate в очереди задач с WS-прогрессом,
target_node/online в BulkOperationItem, диалог выбора целевой ноды и
кнопка в тулбаре инстансов (в пределах одного сервера/кластера)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
При удалении LXC/VM через delete_vm/delete_container чистим таблицу installed_apps
(по server_id+vmid), иначе приложение из каталога оставалось «хвостом» в «Мои приложения».
Шлём WS-эвент appstore_app_deleted для live-обновления списка у владельца.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Добавлена возможность пробросить физический диск ноды в VM по стабильному
пути /dev/disk/by-id/ (scsiN/virtioN/sataN) в карточке «Проброс устройств».
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Карточки на странице «Узлы» дополнены важной информацией:
- версия Proxmox VE (чип «PVE x.y.z»);
- число нод в кластере (для кластерных серверов);
- аптайм узла;
- чипы рабочих областей, в которых состоит сервер.
Бэкенд: эндпоинт /api/{id}/nodes отдаёт version; ответ серверов
дополнен полями cluster_name и workspaces (ServerWorkspaceBrief).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Backend: new proxmox mixins and API routes for firewall, pools, access
control, node admin and disks; extend storage and VM endpoints.
Frontend: node managers (firewall, pool, access, disk, node admin), VM
firewall and hardware tabs, PCI/USB device dialogs, related hooks and
locale strings.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Migrate VMs/LXCs between cluster nodes with live progress tracking
- Configure autostart, boot order, startup/shutdown delays, and delete protection from the instance settings tab
- Group OS templates by detected OS (Ubuntu, AlmaLinux, etc.) instead of a single "Cloud Images" bucket
- Expand native Proxmox backup job management (schedule, compression, retention, guest selection, notifications)
- Add ru/en translations for the above
Add per-user quotas (max instances, vCPU, RAM, disk) enforced at deploy
time and surfaced to users self-service.
Backend:
- UserQuota model (one-to-one, NULL = unlimited) + migration #28
- quota_service: usage from vm_instances cache, check_quota raises 429
- enforce in VM (templates.py) and LXC (lxc.py) deploy paths
- RBAC quota:view / quota:manage, granted to admin
- admin API GET/PUT /admin/api/users/{id}/quota
- self-service GET /settings/api/quota
- tighten VM list isolation via can_view_all_instances
Frontend:
- admin quota dialog in UsersPage with usage bars
- quota card on Settings page (own limits)
- remaining-quota hint in CreateInstanceWizard config step
- useUserQuota/useSetUserQuota, useMyQuota hooks; RU/EN locales
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Backend fixes for VM/LXC disk resize:
- resize_vm_disk/resize_container_disk now honor wait_for_task and raise
the real Proxmox exitstatus (e.g. "disk 'scsi0' does not exist") instead
of always returning success.
- add grow_vm_filesystem: after a VM resize, rescan the disk (retry loop,
needed for SATA/SCSI), growpart the partition and resize the FS
(ext*/xfs/btrfs) via the guest agent; surface its output to the API.
- execute_script normalizes CRLF/CR to LF so pasted scripts and growpart
don't break on trailing \r ("partition-number must be a number").
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- _do_clone_sync now tails the PVE clone task log and maps the disk
transfer percentage to deploy progress 50→95% instead of a static 70%.
- Add owner_id to CloneRequest; admins may assign the clone to another
user (mirrors LXC deploy). Register the cloned instance via
save_vm_instance so it has an owner and is visible to limited users.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replace all legacy dotted permission codes (vms.view, proxmox.manage, ...)
in PermissionChecker/has_permission across app/api with the new
resource:action format (vm:view, server:manage, ...), per LEGACY_PERMISSION_MAP.
All 33 codes used by the API are now registered in the permission registry.
Also switch the /api/auth/me admin branch to get_all_permissions_v2 so the
endpoint emits permissions in the new format only.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Drop dead route handlers that have an active replacement and zero callers
(frontend, backend, tests):
- create-lxc, create-lxc-smart -> superseded by POST /lxc/deploy
- clone-lxc -> superseded by container/{vmid}/clone
- {server_id}/lxc-templates -> duplicate of /lxc-templates/{server_id}
- dashboard/stats -> dashboard uses resources/all + alerts
- permissions (v1) -> superseded by permissions/v2
The identically-named ProxmoxClient/SecurityService methods are kept (they
back the surviving endpoints).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- get_vm_status and get_container_status were missing require_vm_access(),
allowing any user with vms.view to see any VM status regardless of ownership.
Now users with vms.view.own can only access their own VMs (403 otherwise).
- Downgrade upload-artifact@v4 → @v3: v4 is not supported on Gitea (GHES)
and caused CI failures on every run.
Co-Authored-By: Claude Haiku 4.5 <noreply@anthropic.com>
Critical security and robustness fixes from a code audit:
- WS terminal/VNC now require a JWT (?token=), validate the session,
vms.console permission and VM ownership before bridging to Proxmox
(previously the LXC terminal gave unauthenticated root shell access).
Frontend ConsolePage passes the token in both WS URLs.
- Fix SSH fallback for `pvecm delnode`: SSHClient now supports password
auth (it previously crashed with TypeError); ProxmoxClient stores the
password for SSH fallbacks and the host port suffix is stripped.
- Add production startup warnings for default ADMIN_PASSWORD and wildcard
CORS_ORIGINS.
- dashboard: use tz-aware utcnow() instead of naive datetime.utcnow().
- crypto: narrow decrypt_value except to InvalidToken and log failures.
- ssh_client: guard the shared connection cache with a lock; make
unknown-host-key rejection opt-in via SSH_REJECT_UNKNOWN_HOSTS.
- logging: send loguru console output to stderr instead of stdout.
- Sync version fallback to 1.5.1; remove one-off refactor scripts.
- Add CODE_REVIEW.md audit report.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The destroy endpoints returned 400 'Cannot delete a running container.
Stop it first.' when the user clicked Destroy on a running instance.
Since the destroy dialog already requires typing the instance name to
confirm, the action is unambiguously intentional — now both endpoints
issue stop_container/stop_vm with force=True, wait up to 15s for the
status to leave 'running', and only then proceed with deletion.
Previously the reinstall endpoint required cached.template_id (clone-from-VM-
template flow), so LXC instances created via the wizard from an OS template
file (e.g. local:vztmpl/debian-13-...tar.zst) returned 'VM has no associated
template; reinstall is not available'.
Now reinstall works in two modes:
1) clone-from-template (existing): VM/LXC with cached.template_id pointing
to OSTemplate.vmid
2) ostemplate-file (new, LXC only): cached.template_name like
'storage:vztmpl/...tar.zst'. Reinstall stops, deletes, then re-creates
the container with the same VMID using create_lxc_container, preserving
network config, storage, rootfs size, swap, features, ssh_keys, etc.
pulled from the previous container config + VMInstance cache.
Backend:
- Add DeployTask.kind column ('deploy'|'reinstall'|'clone'|'change_password')
+ Migration 23: ALTER TABLE deploy_tasks ADD kind, drop NOT NULL on template_id
- Update DeployTask.to_dict() with kind-aware Russian descriptions
- New backend/app/api/proxmox/async_ops.py with three workers:
_do_reinstall_sync, _do_clone_sync, _do_change_password_sync
All run on shared _deploy_executor thread pool from templates.py and
report progress through _update_deploy_task (DB + WebSocket broadcast)
- Rewrite five vms.py endpoints to return 202 + DeployTaskStartResponse:
POST /vm/{vmid}/clone, /container/{vmid}/clone,
POST /vm/{vmid}/reinstall,
POST /vm/{vmid}/change-password, /container/{vmid}/change-password
Frontend:
- DeployTask: add 'kind' and 'server_id' fields
- useCloneInstance/useReinstallInstance/useChangePassword now return
AsyncTaskResponse {task_id, status, name}; remove invalidateQueries
- InstanceActionDialogs: each onSuccess pushes task to deploy-tasks-store
with appropriate kind, shows 'launched in background' toast
- InstancesPage:
* ghostRows only for kind in {deploy, clone}, deduped by (server_id, vmid)
* reinstall / change_password tasks shown as inline overlay (badge + ring)
on the existing VM row's status column
- AttributeError: OSTemplate has no 'template_vmid' field; the actual column is 'vmid'.
- Use OSTemplate.get_vmid_for_node() to support cross-node replicated templates,
falling back to source_node/vmid otherwise.
- VMInstance.memory is stored in bytes, but Proxmox config expects MB.
Specific endpoints (clone, reinstall, change-password) were being intercepted
by the generic /{action} handler because they were registered AFTER it.
Fix: move CloneRequest, ChangePasswordRequest models, _resolve_server,
_get_client_or_503 helpers, and all 5 specific endpoint functions BEFORE
the generic @router.post('.../{action}') handlers.
Remove duplicate definitions from the old 'Additional Instance Operations'
section at the end of the file.
Root cause of 'Connection closed' / VNCAuth failure:
- vncproxy was called without generate-password=1
- vnc_data.password was None → noVNC sent empty DES-encrypted challenge → Proxmox rejected
Fix:
- Add generate-password=1 to both qemu and lxc vncproxy requests
- If Proxmox returns password field (Proxmox 8.x VNCAuth): use it
- If not (Proxmox 9.x NoAuth): fall back to ticket as password
- Add diagnostic logging (password_len, has_generate_password)
Remove the 'onFirstMessage' hack where backend sent VNC password as
first text WebSocket message and frontend passed the already-open WS
to noVNC's RFBClass. This caused RFB handshake failures because noVNC
couldn't properly manage a WebSocket it didn't create.
Changes:
- Backend: remove send_text(password) before proxy start
- Backend: improve proxy error logging from debug to info level
- Frontend (virtual_machines.html, instance_detail.html): pass WebSocket
URL string to RFBClass constructor - noVNC creates and manages its own
WebSocket connection, allowing proper RFB protocol handshake
- configure_vm: capture resize UPID and call wait_for_task(120s)
before returning; prevents lock conflict when VM is started right
after disk resize
- resize_vm_disk / resize_container_disk (x2): same fix applied
- remove unnecessary restart_vm() after manual QEMU disk resize —
resize is a storage-layer operation and does not require a reboot;
the auto-restart was causing a second lock conflict
Fixes: can't lock file '/var/lock/qemu-server/lock-xxx.conf' - got timeout
Fixes: command '/usr/bin/qemu-img resize' failed: got timeout
- Migration 20: user_servers association table (User ↔ ProxmoxServer)
- API: GET/PUT /api/users/{id}/servers — assign servers to users with workspace-conflict validation
- API: GET /api/users/{id}/server-assignments — enriched server list (assigned + compatible flags)
- Non-admin server filter: intersection of assigned_servers and workspace servers
- Migration 21: dashboard:view = false for user role (new key format)
- Migration 22: proxmox.cluster.manage → cluster:manage for admin role
- API: GET/PUT /api/{server_id}/vm/{vmid}/owner — VM/LXC owner assignment (admin-only)
- Login: smart redirect based on user permissions after authentication
- Login: returnUrl param support for post-login redirect
- Roles API: user_count field per role in GET /api/roles
- node_uptime added to server detail API response
- Remove redundant PermissionChecker from GET /users HTML route
